Figure 8

Sonicated DNA improves the quality of WGA and Illumina Infinium genotyping. (A) This graph shows test:reference product ratios (Y-axis) for PRT assay 2n13 performed using equal amounts of various input DNAs as labelled (X-axis). All PRT reactions were run in quadruplicate, with maximum and minimum plotted as error bars. The dotted horizontal line at 1.43 indicates the ratio that would be produced if the test and reference amplicons amplified with exactly equal efficiency. ‘Control’ indicates that the PRT employed freshly prepared genomic DNA. ‘Intact’ indicates that the same genomic sample was first subjected to WGA using the MDA method (QIAGEN Repli-g Mini kit applied to 50 ng of DNA). ‘Sonicated’ indicates that the same genomic sample was first sonicated to less than 1 kb average size and WGA processed. The blue data points show data produced by the above regimes, whereas the red data points are from an equivalent experiment where the DNA was additionally digested with Nco I immediately prior to inclusion in the PRT reactions. As indicated in Figure 4, Nco I cuts the genomic DNA just upstream of the test sequence target region, and separates it from a nearby region of high C + G content. The data points for the ‘intact’ column demonstrate that after WGA the test locus is still subject to reduced amplification efficiency. Importantly, correction by digestion is substantially reduced compared to the control. Sonication prior to WGA dramatically enhances amplification to almost the efficiency of the references locus even without correction by digestion. (B) Using chromosome 7 as a typical example, log R ratio plots (a measure of relative signal strength) are shown for Illumina Infinium genotyping data generated by assaying a freshly prepared intact genomic DNA sample (log R ratio plot in the upper box) and from a portion of that sample sonicated to 0.3 – 3 kbp in size (log R ratio plot in the second box). The data tracks below these boxes show the apparently reduced signal strength regions (as copy number inferences) generated on the same platform for two poorly performing DNA samples (those mentioned in Figure 1), the C + G content and CpG island maps, and the chromosome 7 ideogram.